Study on Spatial Agglomeration of Services Economy in China and Its Determinants: Spatial Statistical Survey and Econometric Analysis Based on Panel Data of 31 Provinces and Cities

Combining spatial statistics and spatial econometrics, this paper describes and empirically analyzes the status of the spatial agglomeration of services economy in China based on panel data of 31 provinces and cities from 1996 to 2012, and comes up with the following conclusions. First, global Moran's I index is significantly positive which suggests that spatial agglomeration definitely exists in the developing services economy, and the trend of the positive spatial autocorrelation indicates a spiral development. Second, from the scatter diagram of local Moran's I index, Matthew effect characterized – high agglomeration and low–low agglomeration are identified in over 60% provinces and cities and the path dependence characteristic is found in the cluster types of over 90% provinces and cities. Third, using either the 0–1 border adjacency method or the distance attenuation method to measure the spatial weight matrix, the spatial autoregressive coefficient as shown by SAR model and the spatial autocorrelation coefficient as shown by SEM model are significantly positive, further testifying to the existence of the significant positive spatial autocorrelation in inter–provincial services economic development. Moreover, regional market scale, human resource, RD resource and infrastructure can significantly promote the development of local services economy, but industrialization level of the industrial structure would play an inhibiting role.